WHAT WE KNOW ABOUT COMETS 527
(c) 12 or fewer are slightly hyperbolic;
(d) None are strongly hyperbolic.
Now it has been shown by Thraen, Fayet and Fabry in the last two decades that several of the twelve orbits thought to be hyperbolic were not really so, but that they owed their reputations to poor or insufficient observations, or to errors in the computations, and that all of the gen- uine hyperbolas save one acquired their hyperbolicity after the comets concerned came under the disturbing influences of our planets. Five years ago Stromgren was able to show that the one outstanding hyper- bolic orbit was caused, in the same way, by the disturbing attractions of the planets. The original, undisturbed orbit of every one of the so-called hyperbolic comets was, therefore, an ellipse. Fayet has fur- ther shown that a very great majority of the orbits which had been observed to be sensibly parabolic when the comets were near the planets and sun were clearly elliptic when the comets were still far out from the sun; that is, as these comets, moving in elliptic orbits, came in toward (he planets and sun, the attractions of the planets made their orbits approach closely to the parabolic form. There is no reason to doubt that far out in the domain of the sun the comets all approach in elliptic orbits; but that when the attractions of one or more of our planets upon them become appreciable, some of the orbits are changed into shorter ellipses, others are changed into ellipses so long that it is difficult to distinguish them from parabolas, and many orbits are changed to the hyperbolic form. Those comets whose orbits are thus thrown into the hyperbolic form will leave the solar system and travel out through the stellar system.
3. A statistical study of comet orbits made by Leuschner a decade ago bears upon this question. He found that prior to 1756 ninety-nine per cent, of all comets were said to move in parabolic orbits, but that only fifty-four per cent, of comets between 1846 and 1895 were said to move in orbits approximately parabolic; and, secondly, that of comets under observation less than 100 days, sixty-eight per cent, were said to be parabolas, whereas of those observed from eight months to seven- teen months, only thirteen per cent, have orbits approximately para- bolic. These facts point to the conclusion that when comets are ob- served inaccurately, as of old, and in only a short section of their orbits, parabolic orbits satisfy the observations within the limits of the errors unavoidably attaching to those observations; but that when comets are observed accurately and for a long stretch of time, nearly all are found to be moving in ellipses. Most of the ellipses are of course extremely long ones.
If comets starting substantially at rest came from a very great dis- tance away from our sun, say one hundredth the distance of the nearest star, which we think is decidedly within the sphere of our sun's attrac-
�� �
